-
(Random thoughts as I'm struggling with slow ingestion times)
Speaking of multi-threading: further down the line the bottleneck for real-time use cases is probably gonna be the speed of ingestion w…
-
**Is your feature request related to a problem? Please describe.**
I need to use CCL to send neural network weights from one device to another without using host.
Also we need to have all_reduce sup…
-
-
-
Implement Data parallel for Convnet Mnist model on n300.
-
- [ ] Adapt the distilbert data parallel pipeline to function based on the available machine.
-
Particularly when starting with the playlist page the app startup is noticeably slower because playlist were fetched last.
-
-
Thanks for your excellent works！！
I found that if zigzag_ring_attn is used, data_parallel cannot take effect. It seems using all gpu to do seq parallel.
-
### Problem Description
Ran following example:
https://pytorch.org/tutorials/beginner/blitz/data_parallel_tutorial.html with little modification but it failed during run:
if I apply nn.dataParall…